Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
Effect of Vasoactive Intestinal Polypeptide on
Active and Passive Transport in the Human Jejunum
GLENN R. DAVIS, CAROL A. SANTA ANA, STEPHEN G. MORAWSKI, and
JOHN S. FORDTRAN, Department of Medicine, Baylor University Medical Center,
Dallas, Texas 75246
A B S T R A C T The effect of intravenous vasoactive
intestinal polypeptide (VIP) on normal transport mechanisms in the human jejunum in vivo was examined
with the triple-lumen, steady-state perfusion technique. By using special test solutions that revealed
different aspects of jejunal transport, we were able
to evaluate the effect of VIP on specific transport
processes, such as active bicarbonate absorption, active
chloride secretion, and passive absorption or secretion
of sodium chloride. At an infusion rate of 200 pmol/
kg per h, VIP inhibited active bicarbonate absorption
by -42%, stimulated active chloride secretion to a
slight extent, and slightly reduced passive sodium
chloride absorption. A larger dose of VIP, 400 pmol/kg
per h, had essentially the same effect on active bicarbonate absorption and active chloride secretion,
but it markedly depressed passive sodium chloride
absorption and also inhibited passive secretion induced by mannitol. VIP reduced the lumen-to-plasma
unidirectional sodium and chloride flux rates, while
the plasma-to-lumen flux rates were decreased to a
lesser extent or remained unchanged. The potential
difference became more lumen-negative with VIP,
but the sodium diffusion and glucose-stimulated potential were not affected. We conclude that the major
effect of VIP in the human jejunum is to decrease
the normal absorption of water and electrolytes-not
only active bicarbonate-mediated absorption, but also
the passive absorption in response to osmotic forces
generated by active or facilitated absorptive processes.
Although an increase in chloride secretion does occur,
this does not appear to be of major importance.
INTRODUCTION
Previous work has shown that high mucosal concentrations of cyclic AMP stimulate active chloride secretion and inhibit sodium chloride absorption in the
intestine of laboratory animals in vitro (1). Since vasoReceived for publication 15 December 1980 and in revised
form 12 February 1981.
active intestinal polypeptide (VIP)' causes an increase in mucosal concentration of cyclic AMP (2),
it is generally assumed that severe diarrhea seen in
patients with VIP-producing tumors is caused by stimulation of active chloride secretion and by a simultaneous inhibition of normal sodium chloride absorption. It was the purpose of this study to determine
the relative importance of each of these cyclic AMP
effects in man in vivo. This was done by examining
the response of the human jejunum to intravenously
infused VIP. The jejunum was selected for study because it is in this part of the intestine that water
and electrolyte transport is most deranged in patients with diarrhea due to VIP-secreting endocrine
tumors (3).
It is known from previous work that the normal
human jejunum in vivo actively absorbs bicarbonate
via a Na/H exchange (4), and that to a lesser extent
it actively secretes chloride (5). The mucosa is highly
permeable to sodium chloride (6), and passive sodium
chloride absorption or secretion is quantitatively a
major mechanism for salt transport in this part of the
human intestine (7). Therefore, in our studies we paid
particular attention to the effect of VIP on active
bicarbonate absorption, on active chloride secretion,
and on passive absorption and secretion of sodium
chloride.
VIP was infused intravenously at a rate of 200 or 400
pmol/kg per h, since previous studies have shown
that this rate of infusion elevates plasma VIP concentration into the range found in patients with
diarrhea associated with VIP-producing tumors (8).
Use of higher doses is limited by side effects which
include tachycardia, intense flushing, headache, and
fatigue.
METHODS
Subjects. Normal volunteers, aged 21-37 yr, were studied.
Informed written consent was obtained from each subject.
I
Abbreviations used in this paper: PD, potential difference;
PEG, polyethylene glycol; VIP, vasoactive intestinal poly-
peptide.
J. Clin. Invest. © The American Society for Clinical Investigation, Inc.
Volume 67 June 1981 1687-1694
-
0021-9738/81/06/1687/08 $1.00
1687
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
TABLE I
Composition of Test Solutions
Solution
A
B
C
D
E
F
G
Bicarbonatecontaining
Bicarbonatefree
Salinemannitol
Salinefructose
Salinemannitol
Salinemannitol
Salineglucose
mM
Na
K
Cl
HCO3
S04
Mannitol
Fructose
Glucose
Measured
osmolality+±SE
135
5
105
35
135
5
105
17.5
30
125
125
75
50
100
125
125
75
50
100
140
170
300
272
±1
40
40
276
±+1
274
+1
280
+2
274
±1
±
80
272
±
All test solutions contained PEG 2 g/liter.
This project was approved by a Human Research Review
Committee on 1 June 1979.
Jejunal perfusion. Using the triple-lumen tube method,
jejunal perfusion was carried out in the standard manner (9)
when the infusion site was located at the ligament of Treitz.
The mixing segment was 10 cm, and unless otherwise stated,
the test segment was 30 cm, and the test solutions were
perfused at 10 ml/min. The makeup of the test solutions
is shown in Table I. Polyethylene glycol (PEG), 2 g/liter,
was added as the nonabsorbable volume marker. A study
consisted of a 40-min equilibration period, followed by a 60min collection period. Samples ofthe perfusate were aspirated
at a rate of 1.5 ml/min from the proximal and distal ends of the
test segment. Test solutions and aspirated samples were
analyzed for PEG, electrolytes, and solutes by previously
published methods (5). Net water and electrolyte movement was calculated by standard nonabsorbable marker
equations (6).
In some studies (see Results), the test solutions contained
0.5 uCi/liter of 24Na and 36C1, which were assayed in a
Packard 2425 liquid scintillation spectrometer (Packard
Instruments, Downers Grove, Ill.). Unidirectional fluxes were
calculated by the equation of Berger and Steele (10).
In all perfusions, an initial control study, during which
saline was infused intravenously, was followed by a VIP
period during which VIP was infused intravenously. Previous studies have shown that under control conditions
absorption rates are consistent during two consecutive perfusion experiments (11).
VIP. Pure porcine VIP (Viktor Mutt, Karolinska Institute,
Stockholm, Sweden), dissolved in normal saline that contained 0.5% albumin, was infused intravenously at a rate of
200 or 400 pmol/kg per h with a Harvard Apparatus constant infusion pump (Harvard Apparatus Co., Inc., S. Natick,
Mass.).
Potential difference. Potential difference (PD) was measured between the perfusion solution, which served as a
flowing intraluminal electrode, and a subcutaneous reference
electrode (5). Briefly, the reference electrode consisted of
saline in a Medicut catheter (Sherwood Medical Industries,
St. Louis, Mo.) inserted into the subcutaneous tissue of
the dorsal aspect of the forearm. This subcutaneous reference and the flowing intraluminal catheter were connected
1688
via 3 M KCI agar bridges and calomel half-cells to the input
terminals of a battery-powered electrometer (model 602,
Keithley Instruments, Inc., Cleveland, Ohio), and the output
was displayed on a chart recorder (model 281, Rikadenki,
Tokyo, Japan).
RESULTS
Effect of VIP during perfusion of a bicarbonatecontaining solution. When a bicarbonate-containing
balanced electrolyte solution (solution A, Table I) was
perfused in the jejunum, intravenous infusion of VIP
at 200 pmol/kg per h resulted in a significant reduction in water and electrolyte absorption when compared with the control period (Table II). When VIP
was infused at 400 pmol/kg per h, the reduction in
absorption of water and sodium was even greater,
and mean net chloride secretion was observed. Bicarbonate absorption rate was reduced to approximately the same extent by the two doses of VIP;
in both instances, bicarbonate absorption was reduced even though mean luminal bicarbonate concentration was higher during the VIP period than during the control period (due to less bicarbonate absorption in the mixing segment). If bicarbonate
absorption from the test segment is expressed as a
percentage of total bicarbonate which entered the test
segment, VIP reduced bicarbonate absorption by 42 and
51% with the 200 and 400 pmol/kg per h rate of infusion, respectively.2
2 Our
conclusion that VIP inhibits bicarbonate absorption
conflicts with an earlier conclusion from our group (8). This
discrepancy is probably due to the fact that VIP inhibition
of bicarbonate absorption results in higher luminal concentration of this ion during VIP test periods than during control
test periods (Table II). This then mitigates the inhibitory
G. R. Davis, C. A. Santa Ana, S. G. Morawski, and J. S. Fordtran
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
TABLE II
Effect of VIP during Perfusion of a Bicarbonate-containing Solution*
VIP
P1$
Bicarbonate
Mean Nal
Net Nat
Mean Cli
Net Cit
movement
test segment
movement
Mean HCOA
test segment
Net HCO3t
test segment
movement
PD'
mlIhI30 cm
meq/liter
meqIhI30 cm
meqiliter
meq/hl30 cm
meqiliter
meqlh/30 cm
mV
-154
+±18
135.4
+0.6
-19.6
±2.1
126.3
±1.3
-13.0
±2.1
10.5
±0.7
-7.0
±0.4
-1.1
±0.6
-84
±17
138.5
±0.6
-10.2
±2.0
122.5
±1.7
-5.9
±1.9
16.3
±1.2
-4.8
±0.5
-2.2
±0.4
<0.005
<0.025
<0.001
NS
<0.001
<0.02
<0.005
NS
-130
±24
140.2
±0.8
-17.9
±3.4
126.1
±3.1
-10.7
±3.0
16.7
±2.1
-7.8
±0.7
-4.0
±0.8
-26
±21
141.9
±1.1
-3.1
±3.0
121.1
±3.8
+1.4
+2.3
23.2
±3.0
-4.7
±0.9
-6.3
±0.8
<0.001
NS
<0.001
<0.005
<0.001
<0.005
<0.01
<0.025
Net$
movement
VIP (200 pmol/kg/h)
(n = 6)
Control
Chloride
Sodium
Water
VIP (400 pmol/kg/h)
(n = 5)
Control
VIP
P1T
* See Table I for composition.
I (-) net absorption; (+) net secretion.
§ Mean test segment concentration is the arithmetic mean of the concentration at the proximal and distal collecting sites.
1 (-) indicates lumen negative PD.
T Significance of difference determined by Student's paired t test.
solution (P < 0.05 for net water, sodium, chloride,
and bicarbonate secretion and PD).
As can be seen by comparing data in Tables II and
III, VIP had much more effect on water and electrolyte
movement when the jejunum was perfused with an
absorbable balanced electrolyte solution than when it
was perfused with a bicarbonate-free solution.
Effect of VIP on unidirectional fluxes of sodium
and chloride. During perfusion of the bicarbonatecontaining solution, the chloride lumen-to-plasma flux
was reduced with both the 200 and 400 pmol/kg per h
dose of VIP when compared with the control period
(Table IV). The plasma-to-lumen chloride flux was not
significantly changed when either 200 or 400 pmol/kg
per h VIP was infused, though there was a tendency
for it also to decrease with VIP. More or less similar
changes occurred in the sodium unidirectional fluxes:
there was a significant reduction of sodium lumento-plasma flux with 400 pmol/kg per h VIP, while the
plasma-to-lumen sodium flux was not altered to a
statistically significant extent.
During perfusion of the bicarbonate-free solution,
there was a significant reduction in both lumen-toeffect of VIP on further net bicarbonate absorption. In the plasma and plasma-to-lumen chloride fluxes with the
earlier study (8), VIP did inhibit bicarbonate absorption if the
latter is expressed as percent of the total load entering 200 and 400 pmol/kg per h VIP infusion (Table IV).
the test segment (Krejs and Fordtran, personal communication). There was no significant change in the sodium fluxes.
Both doses of VIP caused the PD to become more
lumen-negative, but the difference reached statistical
significance only with the larger dose.
Effect of VIP during perfusion of a bicarbonatefree solution. To evaluate the effect of VIP on active
chloride secretion, it was necessary to perfuse an
electrolyte solution that contained no bicarbonate;
otherwise, normal absorptive processes conceal secretion (5).
During jejunal perfusion of a bicarbonate-free solution (solution B in Table I), water, sodium, chloride,
and bicarbonate were secreted during the control
period, when saline was infused intravenously (Table
III). During intravenous VIP infusion, at either 200
or 400 pmol/kg per h, the jejunal secretion rate of
water and electrolytes was enhanced, and the potential difference became more lumen-negative. Although most of these differences were not statistically
significant with either dose of VIP, when all the data
were combined (regardless of the VIP infusion rate)
it was clear that VIP did enhance secretion rate to a
small extent during perfusion of the bicarbonate-free
Effect of Vasoactive Intestinal Polypeptide
on
Jejunal Transport
1689
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
TABLE III
Effect of VIP during Perfusion of a Bicarbonate-free Solution*
Sodium
Water
Chloride
Bicarbonate
Nett
movement
Mean [NaJ§
test segment
Net Nat
movement
Mean [Cl]§
test segment
Net Clt
movement
Mean [HCOJ§
test segment
Net HCO4t
movement
PlY'
mlIhI3O cm
meq/liter
meqIhI30 cm
meqiliter
meqIhl30 cm
meq/liter
meqIhI30 cm
mV
+12
+6
137.9
±0.9
+3.0
110.0
-2.7
±0.7
1.0
±0.4
+0.7
±0.5
-3.8
±0.7
NS
NS
NS
+2.5
±0.8
+4.8
±1.8
NS
+0.3
±0.2
139.8
±0.8
±0.8
111.4
+0.7
NS
0.5
±0.1
+31
10
±0.8
+5.6
<0.05
NS
NS
Control
+1
±16
139.3
±0.6
+1.5
±2.0
110.0
±0.4
+ 1.0
±2.3
1.4
±2.3
+0.5
±0.5
-3.3
±0.8
VIP
+34
±31
140.4
±0.4
+6.5
±4.4
113.9
+5.8
±1.3
±4.7
3.6
±0.6
+1.7
±0.9
-4.8
±0.4
NS
<0.02
NS
NS
NS
<0.01
<0.05
<0.05
+7
+7
138.5
+0.6
+2.3
±1.0
110.4
+0.5
+1.8
±1.1
0.9
±0.3
+32
140.1
+0.4
+6.0
+2.1
112.5
±13
±0.8
+5.3
+2.2
2.2
±0.5
+0.4
±0.2
+1.4
±0.5
<0.025
<0.05
<0.025
<0.05
<0.05
<0.005
<0.05
VIP (200 pmol/kg/h) (n = 6)
Control
VIP
+
P
±1.8
VIP (400 pmol/kg/h) (n = 5)
P
VIP (200 and 400 pmol/kg/h)
(n = 11)
Control
VIP
P
-3.0
±0.5
-4.2
±0.4
<0.05
See legend of Table II for footnote designation.
We have previously provided evidence of active
chloride secretion by the normal human jejunum
during perfusion of a bicarbonate-free solution (5).
Part of the evidence for this was a discrepancy between the observed chloride flux ratio (L -- P/P -- L)
and the chloride flux ratio that would be expected
if chloride movement were passive (calculated by
Ussing's equation) (12). As shown in Table V, a similar
discrepancy was noted in the present control experiment during perfusion of the bicarbonate-free solution. This discrepancy was also seen during VIP
infusion. Although quantitative inferences must be
interpreted with caution, it is interesting to note
that the difference between the calculated and observed flux ratios (A flux ratio) was significantly
greater during the 400 pmol/kg per h VIP infusion
than during the control period (P < 0.05). (The A
flux ratio was also significantly different when all 11
studies, regardless of VIP dose, were combined
[P < 0.01].) These results would be compatible with
increased active chloride secretion during VIP infusion.
Effect of VIP on fructose and fructose-stimulated
water and electrolyte absorption. Fructose is absorbed from the jejunum by facilitated diffusion by a
1690
mechanism that does not (in contrast to glucose) stimulate active sodium absorption (13). Fructose absorption stimulates the passive absorption of sodium
chloride. To evaluate the effect of VIP on absorption
of fructose and fructose-stimulated water and electrolyte absorption, the jejunum was perfused in
random order with a saline-mannitol solution (solution
C, Table I) and on a separate test day with a salinefructose solution (solution D, Table I). On each test
day the subject received intravenous saline during
control period, followed by intravenous VIP at a rate
of 200 or 400 pmol/kg per h. The difference in absorption of water and electrolytes between the two
solutions when the subject was under the same test
condition (i.e., intravenous saline or VIP) represented fructose-stimulated water and electrolyte absorption.
The saline-mannitol solution was infused at a rate of
10 ml/min, whereas the saline-fructose solution was
infused at a rate of 13 ml/min. This difference in
infusion rates compensated for differences in absorption/secretion rates with fructose and mannitol, so that
similar flow rates would be obtained in the test segment (13), which was 20 cm in this group of studies.
G. R. Davis, C. A. Santa Ana, S. G. Morawski, andJ. S. Fordtran
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
TABLE IV
Effect of VIP on Unidirectional Sodium and Chloride Fluxes
Bicarbonate-free solution*
Bicarbonate-containing solution*
Chloride fluxes
L - Pt
Sodium fluxes
P-. L
L- P
P-. L
Chloride fluxes
L- P
Sodium fluxes
P- L
L- P
P- L
meqIhI3O cm test segment
VIP (200 pmol/kg/h)
(n = 6)
Control
VIP
P5
VIP (400 pmol/kglh)
(n = 5)
Control
VIP
P5
40.3
±4.1
27.2
±2.2
62.8
+2.1
43.2
±2.0
25.0
+2.4
27.6
+2.1
47.1
±4.6
50.0
±4.5
29.8
±4.2
24.0
±2.5
53.2
+4.7
43.0
±3.8
17.9
+2.4
22.7
±3.0
41.3
±5.3
46.9
±5.9
<0.01
NS
NS
NS
<0.025
<0.01
NS
NS
34.4
+5.3
23.8
±3.7
50.1
+6.1
32.2
±3.0
33.3
±2.7
34.5
±3.4
52.9
±2.3
54.4
±2.4
14.7
±1.7
16.1
33.6
+3.8
30.5
+3.5
20.2
26.1
±+1.7
±3.3
±2.6
45.7
±5.2
52.3
±7.1
<0.01
NS
<0.05
NS
<0.005
<0.01
NS
NS
* See Table I for composition.
I L -* P, lumen-to-plasma; P -. L, plasma-to-lumen.
§ Significance of difference determined by Student's paired t test.
TABLE V
Effect of VIP on Chloride Flux Ratios during Perfusion
of a Bicarbonate-free Solution*
Chloride flux ratio
A Flux
200 pmol/kg/h
(n = 6)
Control
VIP
pl
400 pmol/kg/h
(n = 5)
Control
VIP
pII
Observedt
Calculated§
P'
ratiol
0.9±0.1
0.8±0.1
1.1+0.1
1.2±0.1
<0.01
<0.01
0.2+0.1
0.4±0.1
<0. 1
1.0±0.1
0.8±0.2
1.2±0.1
1.3±0.1
<0.025
<0.05
0.2±0.1
0.5±0.2
<0.05
When all 11 studies are combined, the difference between
the A flux ratio during VIP infusion and the control period
is statistically significant (P < 0.01).
* See Table I for composition.
t Observed flux ratio is L -a P flux divided by P -a L flux.
§ Calculated flux ratio for passive ion movement as determined by Ussing equation (12).
"Significance of difference determined by Student's paired
t test.
¶ A flux ratio is the arithmetic difference between calculated
and observed flux ratio.
At 200 pmol/kg per h , VIP did not affect fructose
absorption in six normal subjects; this infusion rate
of VIP slightly decreased mean fructose-stimulated
water and sodium chloride absorption, but this effect
was statistically significant only for sodium (8.1-6.7
meq/h per 20 cm, P < 0.02). When VIP was infused
at 400 pmol/kg per h, there was again no significant
effect on fructose absorption, but fructose-stimulated
water and sodium chloride absorption was markedly
reduced, as shown in Table VI.
Effect of VIP on mannitol-induced secretion. We
next determined to what extent VIP affected mannitolinduced passive secretion. Only the larger VIP infusion rate was studied. As shown in Table VII, when
the jejunum was perfused with a saline solution
containing 140 mM mannitol (solution E, Table I),
there was net secretion of water and sodium chloride
during intravenous infusion of saline. VIP infusion
(400 pmol/kg per h) inhibited this secretion (Table VII).
Effect of VIP on sodium diffusion and on glucosestimulated electrical potential. The jejunum was
perfused sequentially with three isotonic test solutions, two containing either mannitol or glucose
and a low concentration of sodium chloride (solutions F and G in Table I) and another that was a
balanced electrolyte solution (solution A in Table I).
Each solution was perfused for 1 h, at a rate of 10
ml/min. The first 30 min of each perfusion was con-
Effect of Vasoactive Intestinal Polypeptide on Jejunal Transport
1691
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
TABLE VI
Effect of VIP on Fructose and Fructose-stimulated Water and Electrolyte Absorption in Five Normal Subjects
Mean luminal concentration/flow ratet
i.v. Saline
Absorption/secretion rates§
i.v. VIP
(400 pmol/kg/h)
Man*
Fru*
Man
Fru
Fructose
-
20.5
±0.8
-
20.0
±1.5
Water
11.2
±0.2
9.6
±0.2
12.2
±0.4
9.9
±0.2
126.3
±0.5
128.2
±0.8
132.9
128.9
±0.9
133.4
±1.4
Sodium
Chloride
131.8
±0.8
±1.1
i.v. VIP
(400 pmol/kg/h)
i.v. Saline
Man
Fru
,
Man
Fru
A
-4.0
±0.7
4.0
±0.7
P value¶
for A
-4.9
±0.8
4.9
±0.8
-
+27
±9
-70
±16
97
+22
+19
± 11
-8
±6
27
±8
131.4
± 1.4
+3.5
±0.9
-5.6
9.1
±2.2
+3.4
±1.2
+ 1.7
±0.9
1.7
±0.8
<0.025
± 1.8
134.1
±1.7
+3.7
±1.3
-5.2
±1.5
8.9
±2.3
+3.3
±1.5
+ 1.4
±1.0
1.9
±1.3
<0.05
NS
<0.02
* See Table I for composition of solutions. Man, saline-mannitol C; Fru, saline-fructose D.
t Arithmetic mean of concentration (mmol or meq/liter) and flow rates (mllmin) at proximal and distal ends of the 20-cm
test segment.
§ (+) net absorption, (-) net secretion in mmol, ml, or meq/h/20 cm for fructose, water, and electrolytes, respectively.
A is the difference between the transport rates with fructose and mannitol solutions.
¶ Significance of difference determined by Student's paired t test; NS, not significant (P 2 0.1).
sidered an equilibration period, whereas during the
final 30-min period PD was recorded continuously.
The entire experiment was repeated on the same test
day in the same subjects, while VIP (instead of
saline) was infused intravenously at 200 pmol/kg per h.
On a separate test day the entire sequence was
repeated in separate subjects using the 400 pmol/kg
per h rate of infusion of VIP.
During intravenous saline infusion, the PD was
lumen-positive (+5.0±0.6 mV) when the jejunum was
perfused with a saline-mannitol solution, lumen-negative (-3.8±0.4 mV) when perfused with a balanced
electrolyte solution, and even more lumen-negative
(-9.0±0.8 mV) when perfused with a saline-glucose
solution. The PD was slightly more negative with
VIP during perfusion of the balanced electrolyte
solution with both the 200 (-4.7±0.9 mV) and 400
(-4.1±0.4 mV) pmol/kg per h dose than during the
control period. Neither infusion rate of VIP had any
statistically significant effect on the lumen-positive
PD during perfusion of solution F (+6.0±1.3 mV,
+7.0±1.1 mV, respectively, for the low and high dose
of VIP) or on the lumen-negative PD during perfusion of solution G (-10.3±1.4 mV, -10.1±0.9 mV,
respectively, for the low and high dose of VIP).
DISCUS SION
Under normal conditions in man, jejunal absorption
of a balanced physiological electrolyte solution in vivo
is dependent on bicarbonate absorption. Sodium and
bicarbonate are absorbed by an active process involving Na/H exchange (4), and water and sodium chloride
TABLE VII
Effect of VIP on Passive (Mannitol-induced) Secretion
Sodium
Water
Chloride
Net
secretion
Mean [Na]
test segment
Net Na
secretion
Mean [Cl]
test segment
Net Cl
secretion
PD*
mlhI2O cm
meq/liter
meqIhI2O cm
meq/iiter
meqIhl20 cm
mV
88+11
46+10
94.1+1.3
13.4±1.6
97.3±1.6
13.8±2.0
+3.8±0.7
VIP (400 pmol/kg/h)
94.7±1.1
8.3±1.2
97.1±1.1
7.3±1.1
+3.8±0.5
Pt
<0.05
NS
<0.05
NS
<0.05
NS
Control
n = 6.
(+) lumen-positive PD.
t Significance of difference determined by Student's paired t test.
*
1692
G. R. Davis, C. A. Santa Ana, S. G. Morawski, and J. S. Fordtran
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
are then absorbed passively in response to osmotic
pressure and concentration gradients (7). If bicarbonate
absorption is eliminated, by substitution of sulfate for
bicarbonate in the perfused test solution, the jejunum
no longer absorbs water, sodium, or chloride.
The present studies show that VIP inhibits jejunal
bicarbonate absorption, presumably by inhibition of
Na/H exchange. However, it appears unlikely that this
is the only mechanism by which VIP inhibits jejunal
absorption. This is most evident when one compares
the effect of the high and low doses of VIP. Both
doses inhibited bicarbonate absorption by -50%, yet
the high dose inhibited water and sodium chloride
absorption to a much greater extent than the small
dose (Table II). Therefore, the larger dose of VIP
must inhibit absorption by some nonbicarbonatemediated mechanism.
The most obvious possibility, in light of previous
work in animal experiments in vitro (2), was that the
large dose of VIP might induce more active chloride
secretion than the small dose. To evaluate the effect of VIP on active secretion, we perfused the jejunum
with a bicarbonate-free solution, which allows active
chloride secretion to be revealed (5). We found that
VIP does stimulate active chloride secretion (this is
discussed further in the next to last paragraph of this
Discussion); however, the magnitude of this enhanced secretion was rather small, and was not
clearly greater with the large than with the small
dose of VIP. It thus seemed unlikely that stimulation of active secretion played a major role in the
marked inhibition of absorption by the high dose of
VIP during jejunal perfusion of the balanced electrolyte solution.
Since inhibition of active bicarbonate absorption
and stimulation of active chloride secretion appeared
inadequate to explain the observed effect of the large
dose of VIP, it seemed possible that VIP might inhibit passive salt and water absorption. To test for
this, we evaluated the effect of VIP on fructosestimulated absorption. Fructose was chosen for study
because it is rapidly absorbed by the jejunum by a
mechanism that does not enhance active electrolyte
absorption (13); fructose stimulation of water and
electrolyte absorption is mediated passively. Fortunately for our purposes, VIP did not inhibit fructose
absorption rate, so the extent to which fructose absorption was accompanied by absorption of salt and
water could be used to assess the effect of VIP on
passive water and salt absorption. The results revealed that the large dose of VIP markedly inhibited
fructose-stimulated absorption of sodium chloride and
water, whereas the small dose had only a slight
inhibitory effect (Table VI).
The hypothesis that the large dose of VIP inhibits
passive movement of sodium chloride and water was
tested by an additional experiment. When isotonic
solutions containing mannitol are perfused through
the jejunum, the jejunum secretes water and sodium
chloride. This secretion results from passive forces
(6). If VIP inhibits passive absorption of water and
sodium chloride in response to fructose absorption,
it should also inhibit passive secretion of water and
sodium chloride in response to perfusion with mannitol.
The results ofthe experiment supported the hypothesis
(Table VII).
These results offer an explanation for the effect of
VIP on jejunal absorption of physiological test solutions. The small dose inhibits absorption mainly because it inhibits bicarbonate absorption, presumably
via inhibition of Na/H exchange. This dose of VIP also
inhibits passive absorption of water and sodium chloride, and it probably stimulates active chloride secretion, but these effects of the small dose of VIP are
relatively unimportant. The large dose of VIP inhibits
bicarbonate absorption and stimulates active chloride
secretion to about the same extent as the small dose,
but it inhibits passive water and sodium chloride
absorption to a much greater extent than does the
small dose of VIP.
As far as we are aware, no one has previously
suggested that the effect of VIP (or any other disorder of intestinal transport) is mediated in part by
inhibition of passive movement of water and electrolytes. Others have noted that conductance across in
vitro mucosa is reduced by cholera toxin (1); however,
as far as we are aware, it is not known whether
conductance, as measured in vitro, and passive net
water and electrolyte movement, as measured in vivo,
are controlled and determined by the same factors.
How VIP reduces passive movement of water and
sodium chloride is unknown from our studies. VIP
could be reducing passive movement through either
paracellular or transcellular pathways by making these
transport routes functionally more restrictive or by
reducing accessibility to them. Previously it was shown
that there is narrowing of the intercellular spaces during VIP administration to dogs in vivo (14), but whether
this anatomical change represents a mechanism for
functional change is uncertain; it may be the result and
not the cause of alterations in water and electrolyte
transport.
Regardless of the manner by which VIP alters
passive movement, the fact that VIP does not influence
the direction or magnitude of the PD change resulting from perfusion of the jejunum with solutions
containing a low sodium chloride concentration (sodium diffusion potential) suggests that VIP does not
alter the cation-selective nature of the major channels
for passive diffusion of anions and cations. We also
demonstrated that VIP does not change the direction
or magnitude of the glucose-stimulated potential.
Effect of Vasoactive Intestinal Polypeptide on Jejunal Transport
1693
Downloaded from http://www.jci.org on June 14, 2017. https://doi.org/10.1172/JCI110206
This suggests that glucose absorption and glucose
stimulation of sodium movement across the brush
border and lateral membranes are qualitatively intact
during VIP infusion.
During perfusion of either the balanced or the bicarbonate-free solution, the PD was significantly more
negative during VIP infusion. This may be due to a
stimulation of active chloride secretion (assuming
that chloride secretion is electrogenic), since flux ratio
analysis suggested enhanced active chloride secretion
during VIP infusion. It may seem contradictory to
postulate a stimulation of active chloride secretion
when plasma-to-lumen flux of chloride tended to fall
rather than rise (Table IV). However, this apparent
discrepancy is probably explained by the simultaneous inhibitory effect of VIP on passive chloride
movement. In other words, VIP reduces passive
chloride diffusion from plasma-to-lumen (and from
lumen-to-plasma) and increases active plasma-to-lumen
flux; the reduction in passive diffusion is quantitatively greater, so total plasma-to-lumen chloride
flux falls in spite of the stimulated active chloride
secretion.
From our studies we conclude that the major
effect of VIP in the human jejunum in vivo is to
decrease normal absorption of water and electrolytesnot only active bicarbonate absorption, but also passive
absorption of water and sodium chloride in response
to osmotic forces generated by active or facilitated
solute transport. Although an increase in active
chloride secretion does occur, this does not appear
to be of major importance, at least not with these
doses of VIP. It is possible that larger doses might
induce an important, or even dominant, active electrolyte secretion by the jejunum. However, it should
be noted that the infusion rates of VIP that we employed result in steady-state plasma VIP concentrations which are well within the range of values
found in patients with severe secretory diarrhea apparently caused by VIP-secreting endocrine tumors (8).
So, the doses of VIP which we infused are in
the pathophysiological range of interest, even though
they do not constitute a full dose-response study.
Studies with significantly larger doses of VIP (in
terms of a dose-response curve) will probably never
be possible in humans because they result in intolerable side effects.
ACKNOWLEDGMENTS
The authors express their appreciation to Ms. Jean Harber
for her excellent assistance in preparation of this manuscript.
1694
This work was supported by the Department of Health,
Education and Welfare Research grant 1-ROl-AM26794
from the National Institute of Arthritis, Metabolism and
Digestive Diseases, and the Southwestern Medical Foundation's Abbie K. Dreyfuss Fund, Dallas, Tex.
REFERENCES
1. Field, M. 1971. Ion transport in rabbit ileal mucosa.
II. Effect of cyclic 3',5'-AMP. Am. J. Physiol. 221:
992-997.
2. Schwartz, C. J., D. V. Kimberg, H. E. Sheerin, M. Field,
and S. I. Said. 1974. Vasoactive intestinal peptide stimulation of adenylate cyclase and active electrolyte
secretion in intestinal mucosa. J. Clin. Invest. 54: 536544.
3. Krejs, G. J., J. H. Walsh, S. G. Morawski, and J. S.
Fordtran. 1977. Intractable diarrhea: intestinal perfusion
studies and plasma VIP concentrations in patients with
pancreatic cholera syndrome and surreptitious ingestion
of laxatives and diuretics. Am. J. Dig. Dis. 22: 280-292.
4. Turnberg, L. A., J. S. Fordtran, N. W. Carter, and F. C.
Rector. 1970. Mechanism of bicarbonate absorption and
its relationship to sodium transport in the human jejunum.
J. Clin. Invest. 49: 548-556.
5. Davis, G. R., C. A. Santa Ana, S. G. Morawski, and J. S.
Fordtran. 1980. Active chloride secretion in the normal
human jejunum. J. Clin. Invest. 66: 1326-1333.
6. Fordtran, J. S., F. C. Rector, M. F. Ewton, N. Soter,
and J. Kinney. 1965. Permeability characteristics of the
human small intestine. J. Clin. Invest. 44: 1935-1944.
7. Fordtran, J. S., F. C. Rector, and N. W. Carter. 1968.
The mechanism of sodium transport in the human
small intestine.J. Clin. Invest. 47: 884-900.
8. Krejs, G. J., and J. S. Fordtran. 1980. Effect of VIP infusion on water and ion transport in the human jejunum.
Gastroenterology. 78: 722-727.
9. Cooper, H., R. Levitan, J. S. Fordtran, and F. J. Ingelfinger. 1966. A method for studying absorption of water
and solute from the human small intestine. Gastroenterology. 50: 1-7.
10. Berger, E. Y., and J. M. Steele. 1958. The calculation
of transfer rates in two compartment systems not in
dynamic equilibration J. Gen. Physiol. 41: 1135-1152.
11. Gray, T. K., P. Brannan, D. Juan, S. G. Morawski,
and J. S. Fordtran. 1976. Ion transport changes during
calcitonin-induced intestinal secretion in man. Gastroenterology. 71: 392-398.
12. Ussing, H. H. 1949. The distinction by means of tracers
between active transport and diffusion. The transfer of
iodide across the isolated frog skin. Acta Physiol. Scand.
19: 43-56.
13. Fordtran, J. S. 1975. Stimulation of active and passive
sodium absorption by sugars in the human jejunum.
J. Clin. Invest. 55: 728-737.
14. Krejs, G. J., R. M. Barkley, N. W. Read, and J. S.
Fordtran. 1978. Intestinal secretion induced by vasoactive intestinal polypeptide. A comparison with cholera
toxin in the canine jejunum in vivo. J. Clin. Invest.
61: 1337-1345.
C. R. Davis, C. A. Santa Ana, S. G. Morawski, and J. S. Fordtran